Methodology for predicting brake squeal propensity using complex eigenvalue analysis, including thermo-mechanical effects.

With brake squeal being the most prevalent noise vibration and harshness issue in modern vehicles, this paper presents an improved methodology for brake squeal propensity prediction at the design stage. The research established four clearly defined 'Stages' in conducting finite element squeal analyses, describing crucial input data, modelling procedures, output and validation results. Stage 1 deals with free-free modal characteristics of individual brake components and their material characteristics. Stage 2 combines individual parts, conducting brake assembly mechanical finite element analyses. Stage 3 concentrates on fully coupled thermo-mechanical finite element analyses, and the concluding stage, Stage 4, focuses on brake assembly stability analyses. Validations proved that very accurate predictions are possible, but the geometries, material characteristics and established modelling procedures must be strictly followed. Material characteristics were most prone to introduce discrepancies with measured values. 'Generic' values are found to be unacceptable and conducting own measurements was necessary, in particular for the friction material, whose anisotropic properties have been measured in detail, leading to high accuracy in predicting pad natural modes and frequencies. In Stage 4, the stability analyses of the full brake assembly were based on the complex eigenvalue analysis (which included thermal aspects), with the sign of the real part giving an indication of stability and the imaginary part defining the frequency of the unstable mode. Instabilities and frequencies predicted match well with the values measured in dynamometer tests, clearly demonstrating the influence of thermal effects. The final output of the procedures described in this paper is a validated three-dimensional thermo-mechanical finite element noise vibration and harshness brake assembly model in which natural frequencies and modes, instabilities and contributing factors can be predicted at any time during a brake application. [ABSTRACT FROM AUTHOR]